Journal of General Microbiology (rg70), 64,255-268
Printed in Great Britain
255
Mouse Virulence of Salmonella Strains: The Effect of
Different Smooth-type 0 Side-chains
By V. V. V A L T O N E N
Department of Serology and Bacteriology, University of Helsinki and
the Central Public Health Laboratory (State Serum Institute),
Helsinki, Finland
(Accepted for publication 16 October 1970)
SUMMARY
A loss or reduction in the 0 side-chain material of the cell-wall lipopolysaccharide is known to abolish or reduce the virulence of Salmonella
strains. The present report concerns the effect on virulence of altering the
quality of smooth-type 0 side-chains in a basically virulent Salmonella typhimurium line. The original rfbB locus determining the structure of 4, IZspecific repeating units was replaced either in transduction or in conjugation
by the wild-type rfbD locus of group D (0 antigens 9, IZ), or rfbc of group C
(0 antigens 6, 7). The LD50 values of the 4, 5, 12 recombinants or transductants were about 105 and like those of the 4, 5, 12 parent, whereas the
LD 50 values of the 9,12 transductants were about I O ~ ,and the LD 50 values
of the 6, 7 recombinants and transductants were over 10'. The reduced
virulence of both 9, 12 and 6, 7 recombinants could be restored to the
original level by reintroducing the rfbB locus into these strains through
conjugation. It seems, therefore, that different kinds of 0 side-chains confer
different degrees of virulence on S. typhimurium.
INTRODUCTION
It has long been known that the change in colony morphology from smooth to
rough (S -+ R variation), with accompanying loss of 0 agglutinating ability, greatly
reduces virulence of Salmonella bacteria (Lingelsheim, I g I 3 ; Arkwright, 1927).
Since the first observation in 1913, further evidence has accumulated for the importance of the structure of the 0 antigen in virulence of Salmonella strains (Roantree,
1967).
The 0 antigen - also called endotoxin or lipopolysaccharide (LPS) - forms the
outermost layer in the Salmonella cell wall and contains a lipid and a polysaccharide.
The latter has a core which is common to all Salmonella species, and to which are
attached long 0 side-chains (Luderitz, Jann & Wheat, 1968). The 0 side-chains of at
least groups B, D and E are polymers of several repeating units per side-chain. The
sugar composition of the repeating units (the 0 side-chain) determines the 0 specificity
characteristic for each Salmonella group (Fig. I).
The S -+ R variation is caused by mutations blocking the synthesis of the polysaccharide. The avirulent R forms thus have an incomplete LPS with no 0 side-chains;
rfa mutants are blocked in the synthesis of the core, rfb in the synthesis of the repeating
unit, and rfc in their polymerization (Makela & Stocker, 1969). Both rfa and rfb
Vol. 64, No.
2
was issued
I
April 1971
17-2
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Thu, 13 Jul 2017 00:41:57
256
V. V. V A L T O N E N
mutants are phenotypically R, but rfc mutants have one 0-specific repeating unit
per side-chain and are SR, intermediate between S and R. Genetic crosses between
different 0 antigen groups (Makela, 1966) have demonstrated that the rfb cluster of
genes, which is close to the his locus on the bacterial chromosome, contains all the
information required for the synthesis of 0 repeating units. If a Salmonella has the
rfbB (of group B), its LPS has the group B specificity 4,12, and if the rfb is of group D
(rfbD)the bacterium has the group D specificity 9, 12, etc.
Group of Salmonella and the
0 antigenic specificity
Structure of 0 side-chain
I
(:::-+ )
/Abe
\
i d a n + Rha-t Gal n
Group B (4, 12)
Rha-t Gal n
Glu
Group D (9, 12)
)
(Man+ M a z a n - + Man-, GNAc n?
Group
c (6,7)
Fig. I . Schematic structure of the 0 side-chains. According to Liideritz, Jann & Wheat
(1968); the structure for the group C side-chain is only tentative (Fuller & Staub, 1968).
Abe = abequose, Gal = galactose, Glu = glucose, GNAc = N-acetylglucosamine,Man =
mannose, Rha = rhamnose, Tyve = tyvelose.
Some recent studies suggest that even in non-R forms possessing 0 side-chains the
quantity of the 0 side-chain material plays a role in virulence. The mouse virulence
of an SR form, which has only one repeating unit per side-chain, is intermediate
between the virulence of the S and that of the R form (Nakano & Saito, 1969). A highly
virulent smooth Salmonella typhimurium was shown to possess about twice as many
of the antigenic determinants 4, 5 and 12, measured by an immunological technique,
as another smooth but avirulent S. typhimurium strain (Archer & Rowley, 1969).
The ribose- and galactose-containing T I side-chains appear to confer only a slight
degree of virulence on a rfb mutant of S. typhimurium, much less than do the original
4, 5, 12 side-chains, suggesting that the quality of the side-chain may also affect
virulence (Valtonen, 1969).
The purpose of this work was to make a systematic study of the possible effects
on virulence of altering the quality of smooth-type 0 side-chains in a known virulent
line of Salmonella typhimurium. The alteration was accomplished by genetic replacement of the original rfbB locus (that is rfb of group B), which determines the structure
of 4,ia-specific repeating units, with the rfbD locus of group D (0 antigens 9, 12) or
with rfbc locus of group C (0 antigens 6, 7). The LD50 value for mice was used as
a measure of virulence in this work.
METHODS
Bacterial and phage strains. The bacterial strains with their main properties and
origins are listed in Table I. The basic strain was SL 1027 of Salmonella typhimurium
line L T ~ .This strain was originally obtained from Dr B. A. D. Stocker, Stanford
University, California, U.S.A. The SH strains are kept in the collection of Professor
P. H. Makela, Central Public Health Laboratory, Helsinki, Finland. New mutations
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Thu, 13 Jul 2017 00:41:57
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Thu, 13 Jul 2017 00:41:57
SH1331, I333
~ ~ 1 3 61361
0,
SH I 605
~~2183
SH2201, 2202, 2203
~ ~ 2 2 0 2205,
4 , 2206
SH 2233
~ ~ 2 2 4 2242
1,
SH2245, 2246
SH 2259
~ ~ 2 2 6 2267
6,
~~2268
SH2273, 2274
SH 2275
~ ~ 2 7 1 0
SW I403
SH I
SH 101 I
SH 1036
SL 748
SL go I
SL I 027
SL I 060
Strain no.
FFFFHfrHz
FFFFFHfrH I 4
FFFFFFFFFFFF-
B
B
B
B
B
B
C
D
B
C
C
B
B
D
D
C
B
C
B
D
B
C
B
rfb
LPS
from
characteristic group
-
leu-1357
his-4114 metAaa trpB2 str jlaA66
metAz2 trpB2 str jlaA66
metAa2 trpB2 str jlaA66
his-4118 metAza trpBa str jlaA66
metAaa trpBz str jlaA66
metAaa trpB2 str jlaA66
his-4120 metAaa trpBz str
metA2t trpB2 str jlaA64
metA2z trpBz str flaA66
metAza trpB2 str jlaA66
metAa2 trpB2 str jlaA66
-
metAzz trpBa str flaA66 rfb-430
metAzz trpBa str jlaA66 rfc-497
metAz2 trpBa str jlaA66
metAaa trpB2 str ~7aA66rfaH487
m e t - I I ~ aro-851
I
str-501
hislF135
str purGgot
str purGjo2
-
Genetic markers 11
B. A. D. Stccker; Gemski & Stccker (1967)
B. A. D. Stocker; Gemski & Stccker (1967)
B. A. D. Stocker; Gemski & Stocker (1967)
B. A. D. Stocker; Wilkinson & Stocker (1968)
Makela (1963)
P. E. Hartman; Loper et al. (1964)
P. H . Makela
P. H. Makela
his+ transductants from SH IOII x SH I $
his+ transductants from SH IOII x SH I
P. H. Makela; Makela (1966)
SL 1 027 (dES)$
his+ transductants from SH 1036 x ~ ~ 2 1 8 3
his+ transductants from SH 1036 x ~ ~ 2 1 8 3
SH 2206 (dES)
his+ recombinants from SH 1605 x ~ ~ 2 1 8 3
his+ recombinants from SH I 605 x SH 2 I 83
SH2241 (dES)
his+ recombinants from sw 1403 x ~ ~ 2 2 3 3
his+ recombinants from sw 1403 x SH 2233
his+ recombinants from sw 1403 x ~ ~ 2 2 5 9
his+ recombinants from sw 1403 x ~ ~ 2 2 5 9
Fresh isolate from a patient
Source and reference
p Strain with the rfb but not his derived in a conjugational cross from a strain of
Strain with the rfb but not his derived in a conjugational cross from a strain of S. montevideo (group C ) as donor.
S. enteritidis (group D ) as donor.
$ Transductions (by phage P22) and crosses are described as: (donor) x (recipient).
0 Mutagenesis by diethylsulphate.
11 Gene symbols as follows: uro = phenylalanine+ tyrosine biosynthesis; fla = flagellar synthesis; gal = galactose utilization; his = histidine biosynthesis;
met = methionine biosynthesis ;pur = puririe biosynthesis ; rfa, rfb, rfc = biosynthesis of the 0 antigenic lipopolysaccharide; str = streptomycin resistance;
trp = tryptophan biosynthesis; xyl = xylose utilization.
*
S .typhimurium B
S. typhimurium B
S. typhimurium B
S. typhimurium B
S. typhimurium B
S. typhimurium B
S. typhimurium B
S. typhimurium B
S. typhimurium B
S. typhimurium B
S. typhimurium B
S. abony B
S. typhimurium B
S. typhimurium* B
S. typhimuriumt B
S. typhimurium B
S. typhimurium B
S. montevideo C
S. typhimurium B
S. typhimurium B
S. typhimurium B
S. typhimurium B
S. typhimurium B
Species and group
Mating
type
Table I. Salmonella strains used
258
V. V. VALTONEN
were induced by adding diethylsulphate (10 mg./ml.) to an overnight broth culture
and incubating for 30 min. at 37". The culture was then diluted I :I 00 in fresh medium
and incubated overnight before plating on appropriate suboptimally supplemented
media. The genetic donors were Hfr derivatives of Salmonella abony (Makela, 1963)
and S. montevideo. The gene symbols are explained in a footnote to Table I.
The S- and R-specific phages were originally obtained from Dr R. G. Wilkinson,
University of Melbourne, Parkville, Australia. The following phages were used in this
work: S-specific P22c2, g NA, P22h; R-specific P221, Br2, Ffm, Br60, CZI, S 13,
#X174 and 6SR; and FO which attacks both S and R bacteria (Wilkinson, 1966;
Wilkinson & Stocker, 1968). Drops of phage lysates containing about 108 plaqueforming units were applied to an agar plate previously spread with the bacterial
strain to be tested.
rfb his
Fig. 2. A simplified chromosomemap of Salmonella (according to Sanderson, 1967; Makela
& Stocker, 1969). Positions not accurately known are indicated by symbols in parenthesis.
The point of entry and direction of injection of donor chromosomes are indicated by
arrows on the circle. For gene symbols, see footnote to Table I .
Phage P 22 was used for transduction experiments.
Media. Broth: Difco antibiotic medium 3 (1'75 g./Ioo ml.). Nutrient agar: broth
solidified with 1-3g. agar/Ioo ml. Selective media: Davis minimal medium (Lederberg,
1950) containing agar, 1-2%; glucose, 0.2 %; required amino acids, 20 pg./ml. ;
streptomycin, I mg./ml. where indicated. Suboptimally supplemented media : selective media with only 0-2pg./ml. of the amino acid for which mutants were selected
(Hartman, Loper & Serman, 1960).
Genetic methods. In conjugation experiments, equal amounts of exponentially
growing donor (Hfr) and recipient (F-) broth cultures, containing about 5 x 108bacteria/ml., were mixed and incubated at 37" for 2 h. without shaking before plating
on appropriate selective media (Makela, I 966). In transduction experiments, the phage
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Thu, 13 Jul 2017 00:41:57
Virulence of Salmonella and 0 antigen
259
( 109 particles/ml.) was mixed with the exponentially growing recipient (109 bacteria/
ml. broth), and the mixture left for om in. at 37" with gentle shaking and then
plated on the selective medium (Hartman et al. 1960). The recombinants or the transductants appearing after incubation for 48 h. were restreaked on nutrient agar plates
from which single colonies were picked for testing. The Hfr strains used and a schematic map of the Salmonella chromosome are given in Fig. 2.
Serological methods. 0 and H antigens were determined by slide agglutination in 4 %
saline and in antisera appropriately diluted with 0.2 % saline. Anti-0 sera were prepared by immunizing rabbits with saline suspensionsof nutrient agar cultures of smooth
strains kept at 100' for 2 3 h. ;for anti-H sera the immunogen was a motile broth culture
killed with formalin (Kauffmann, 1966).
Virulence test. The LD50 value in mice was used to measure the virulence of the
Salmonella strains (see Statistical methods). The SAW mice (Swiss Albino Webster
strain) used in this work came from two different sources; line I from the State Serum
Institute, Helsinki, Finland, and line I1 from the Department of Serology and Bacteriology, University of Helsinki. There was a small difference in their sensitivity to
Salmonella typhimurium infection (see Results). Line I was used in only one set of
experiments. Test animals of both sexes (aged 8 to 10 weeks, weighing between 22
and 25 g.) were used.
The bacterial strain to be tested was grown in broth for 18 h. (to early stationary
phase, about 109 bacteria/ml.), tenfold dilutions of the culture were made in 0.9 %
saline and the viable count was determined. Five consecutive dilutions (0.5 ml. each)
were injected intraperitoneally into five groups of ten mice kept in separate cages.
Special care was taken to keep the cages physically separate and to prevent contamination in feeding and handling. The animal room was disinfected after each experiment.
Statistical methods. The LD 50 value was calculated from 10 day survival according
to Reed & Muench (1938). In the standard assay, virulence of two recombinant classes
from one cross (or transduction) possessing different 0 antigens was compared. The
statistical significance was calculated by forming a set of 2 x 2 tables from original
data, according to the method of Mantel & Haenszel (1959). As a rule, three such
tables for three different dose levels were made as follows:
Table of the iih.dose level
No. of mice
Recombinant class I
Surviving
Dead
Xli
Sum
where
Recombinant class I1
N1i - xli
Nl i
Total
xi
Ni - xi
NI
xli and xzi= numbers of surviving mice which had received the same
dose (a) of recombinant class I bacteria (x1J or recombinant class I1 bacteria ( x , ~ ) ,
NIiand N2i= total numbers of mice inoculated with the same dose (a)
of recombinant class I bacteria (Nli) or recombinant
class I1 bacteria (N2i),
xi = xli+xZi(= total number of surviving mice),
Ni = Nli Nzi ( = total number of infected mice).
+
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Thu, 13 Jul 2017 00:41:57
V. V. VALTONEN
260
If we consider the distribution of, say, xll conditionally on the marginal frequencies
Nll, NZ1,xi, N i - x i , under H,, (null hypothesis) it has the hypergeometric distribution a. Under these circumstances x2 should be computed as the ratio of the square of
a deviation from its expected value to its variance (Mantel & Haenszel, 1959).
Since,
E(X11) = XiNlI/N,
and
var(x1,) = NJVZixi(N1- xi)/Ni2(Ni- I),
then
has a x2 distribution with one degree of freedom. The probability that a stated value
of x2 would be exceeded under H,, is tabulated as P.The result is considered significant,
when P < 0.01and highly significant with P < 0-001.
An example of the calculation from the original data in Table 2 is as follows :
Dose level (5 x
Surviving
Dead
Sum
104)
28
2
50
30
60
25
35
30
5
30
60
2
28
9
I1
21
30
30
49
60
22
8
30
Dose level (5 x 105)
Surviving
Dead
Sum
I0
20
Dose level (5 x I O ~ )
Surviving
Dead
Sum
(P <
I0
25
0.001)
Table 2. The lethal eflect in mice of smooth 4, 5, 12 or
smooth 9, 12 recombinants of Salmonella typhimuriurn
The table records the number of deaths in groups of
of overnight broth cultures of S (4, 5, 12) or S (9,
10mice infected with
12)
tenfold dilutions
recombinants of S. typhimurium.
Dose (bacterialmouse)
Strain
s
s
s
SH220I (4, 5, 12)
SH2202 (4, 5, 12)
SH2203 (4, 5, 12)
~ ~ 2 2S
0 (49 , 1 2 )
~~220
S 5(9,12)
~~22S
0 (69 , 1 2 )
5 x 107*
I0
I0
I0
8
8
I0
5 x 1o6
5 x 105
5 x 104
I0
5
5
8
I0
7
5
9
6
9
I
2
2
I
2
0
2
0
5x
103
0
0
0
0
0
0
* For ~ ~ 2 2 the
0 1 actual doses were 6 x IO', 6 x I O ~ etc.,
,
and for ~ ~ 2 2 0the3 corresponding doses
were 4 x 107,4 x I O ~ etc.
,
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Thu, 13 Jul 2017 00:41:57
Virulence of Salmonella and 0 antigen
26 I
RESULTS
Reliability of the virulence test
Elimination of cross-infection. Control mice (I 10 in all) were injected intraperitoneally with 0.5 ml. of 0.9 % NaCl. Most of them were kept in their own cages
placed between experimental cages, but a small number (15) were kept in groups of
three in the same cages as infected animals. None of the control mice died during the
1 0 day observation period.
Stability of the strain during virulence test. Bacterial stability was controlled in each
experiment by testing 20 colonies for their 0 antigen serologically and with S- and
R-specific phages: 20 colonies were obtained both from the inoculum and from the
cultures isolated from heart blood and intraperitoneal fluid of several dead mice.
Most isolations were made within I h. from the moment of death. No reversions from
R to S forms or other changes in the 0 antigens were found after animal passage.
Table 3. Time of death in mice after infection with virulent and
avirulent strains of Salmonella typhimurium
Cumulativedeaths in groups of ten mice injected intraperitonellywith 0.5 ml./mouseof tenfold
dilutions of overnight broth cultures of virulent and avirulent strains of Salmonella
typhimurium.
Number of mice dead after injection with
Days
after
injection
0
SH2I83 s (4,5912)
Dose (bacterialmouse)
A
f
h
r
l
5 X I O 7 5 X I 0 6 5 X I 0 5 ~ x 1 0 4 ~ x 1 50XsI 0 5 ~ x 1
0
0
0
0
0
0
0
I
0
4
9
9
9
I
I
I
0
0
0
0
2
2
I
I
0
0
6
8
3
1
0
0
O
I0
I0
5
I0
I0
I0
5
6
4
4
I0
7
5
5
6
LD50 = I x
2
2
5
105
<R>
SH 2710 s (4,5912)
Dose (bacteria/mouse)
0
0
0
0
J
10
I0
I0
1
1
0
0
I
O
1
0
I0
&
\
0 4 ~ x ~ ~ s 50~ x~ XI Ioo z8 5 X I o 57 x 1 0 ~
0
0
0
3
5
5
g
9
g
g
SL748
Dose(bacteria/mouse)
c
0
0
0
0
0
0
0
0
I
I
0
0
0
8
6
7
7
9
9
9
9
9
0
3
3
3
3
5
5
7
0
2
I
2
0
0
0
0
0
0
0
0
0
0
0
I
I
3
5
5
6
8
I0
I0
I0
I
LD50 < I O ~
The LD5o value calculated from the
10day
8
9
I0
I0
I0
I0
I0
10
I0
I0
3
3
3
5
5
5
5
L
2
7
LD5O = g X
10'
survival.
Choice of the observation period. For obvious reasons a relatively short observation
period was desirable. Essential information does not seem to be lost by our choice of
the 10day period as compared with the conventional 30 day experiment, because over
go % of the deaths in 30 day experiments occur during the first 10 days (Table 3).
The shorter period may have an advantage in recording deaths due to the original
inoculum since it is not likely to include deaths from secondary infection from cage
mates.
Choice of the Salmonella strain. We wanted to use Salmonella typhimurium L T 2 line
as a basic strain, because it is genetically well characterized. The problem was that the
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Thu, 13 Jul 2017 00:41:57
V. V. V A L T O N E N
262
LT 2 line has a fairly low virulence compared with many S. typhimurium strains isolated
1 0
Table 3, LD50 < 102. The LD50 values of
from clinical specimens, e.g. ~ ~ 2 7 in
SH I and SL 1027 used in this work were about 105, while several other smooth LT 2
lines were even less virulent (LD5o > 10~).
New auxotrophic mutants isolated from
SL 1027 after mutagen treatment had often lost their virulence, and for this reason the
virulence of each new mutant was checked before use. In further tests some of the
new avirulent mutants were still smooth, but their loss of virulence could not be clearly
correlated with the known auxotrophic mutations because their prototrophic revertants were still avirulent. The same phenomenon has been described earlier (Herzberg,
1962).
Reproducibility of the LD50 values in the same mouse line was good (Table 4).
Table 4. LD50 values for smooth Salmonella typhimurium strain
and its SR or R mutants
Strain
no.
Somatic
antigen
~ ~ 1 0 2S
7 (4,512)
S L ~ O I SR (4, 5 1 2 )
SL 748
R
SL I 060
R
Mutant
LPS
genes
rfc
rfb
rfaH
SAW
mice line I
SAW
r
(0
4 x IO*
1 x I06
5 x 10'
(ii)
3 x 104
2 x 108
1x10~
zx
I X 108
108
(iii)
I X 106
-
1x10~
-
L T ~
mice line I1
A
(iv)
5x10~
-
-
\
(v)
1x10~
-
(i) etc. refer to different experiments.
Efect of S + R variation on virulence in the chosen experimental conditions. The LD 50
value of the smooth LT2 line was about 105, while the LD50 values of rough forms
(rfb, rfaH) were about I O (Table
~
4). The difference between S and R forms, therefore,
was only about a thousandfold as compared to the almost millionfold difference seen
between S and R forms when starting from highly virulent Salmonella strains (ArkWright, 1927). However, in spite of this narrow scale, the virulence of the semirough
mutant (rfc) with only one repeating unit per side-chain was clearly intermediate
(LD50 1 0 ~between
)
the virulence of the smooth and the R forms. There was no
difference between the rfb and the rfaH mutants, agreeing with previous observations
on different R mutants (Nakano & Saito, 1969; Edebo & Normann, 1970). Although
both mutants were avirulent, they could be isolated as pure cultures from the heart
blood of dead mice, indicating that when injected in large enough amounts they were
also capable of causing infections. The endotoxin (LPS) effect seems to be lethal only
at very high dose levels (> 1o8 bacteria/mouse): some deaths were observed on the
first day after intraperitoneal injection of about I o9 heat-killed bacteria/mouse, but
there were no deaths after a dose of about I O ~ .
-
The efect of 9, rz-speci~kside-chains
The rfb locus of Salmonella typhimurium determining 4, 12 specificity was replaced
through P22 transduction by the rfb locus of S. enteritidis determining 9,12 specificity.
Phage P 22 grown on the his+ 9,12 donor SH 1036 was applied to the his- #,5,12viru3 ; of the his+ transductants had the 0 antilent (LD5o = 105) recipient ~ ~ 2 1 8some
gens 9, 12 indicating that they had inherited the rfbD locus of the donor, while the
remaining his+ transductants were antigenically 4, 5, 12. The virulence of three
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Thu, 13 Jul 2017 00:41:57
Virulence of Salmonella and 0 antigen
263
transductants from both classes was determined (Tables 2, 5). The LD50 values of
three his+ 4, 5, 12 transductants were about 105,like that of the recipient, while the
9, 12 transductants showed slightly reduced virulence (LD50 10~).The difference
between these two recombinant classes was highly significant (P < 0.001).As a con-
-
trol, a his- mutant was made from one 9 , 1 a transductant and was used as a recipient
in a conjugational cross with an avirulent (LD50 = I O ~ )4 , 5 , 12 S.abony Hfr strain.
It was very unlikely that the avirulence loci in S. abony (Krishnapillai & Baron, 1964)
would have been transferred in this conjugation, because of the his+ selection and the
direction of chromosome injection by the donor (Fig. 2). The LD50 values of two
his+ 4, 5, 12 recombinants obtained in this cross were about 1oS,indicating a restoration of the original virulence together with the 4, 5, 12 specificity. The virulence of
a his+ 9 , 1 2 recombinant was unchanged (LD50 > 10~).It thus appears that 4 , 5 , 1 2 specific side-chains confer a slightly higher degree of mouse virulence on a S. typhimurium strain than do 9, la-specific side-chains.
Table 5.LD50 values of smooth Salmonella typhimurium strains
with the 0 specijicity 4, 5, 12 or 9, 12
Origin of transductants or recombinants
L
r
Donor
Recipient
>
P22 transduction
A
I
~~1036x
(his+, 9,12)
LD 50 values
Strain no.
0 specificity
Recipient
>
~ ~ 2 1 8 3
(his-, 4 , 5 , 1 2 )
Selection for his+
Back-cross (conjugation)
~ ~ 2 1 8 3
Transductants
SH 2201
SH 2202
SH2203
SH2204
SH2205
SH2206
his+, 4 , 5 ,
12
X
SH2233
h i s , 9,12
Selection for his+
-
3 x 106
3 x I06
Ix
106
Recipient
SH2233
SW1403
9912
Recombinants
~~2266
SH2267
SH2268
2 x 106
-
-
4 x 106
SH 2233 (his-, 9,12) was derived from transductant SH 2206.
The efect of 6, 7-specijic side-chains
To convert the type 4, 5, 12 Salmonella typhimurium to type 6, 7, conjugation was
first used instead of transduction, because there were no suitable transducing phages
available for such an intergroup system. The results were later confirmed with 6, 7
and 4, 5, 12 transductants, obtained in the round about way described below.
3 ) before
The same virulent his- 4 , 5 , 1 2 Salmonella typhimurium recipient ( ~ ~ 2 1 8 as
was crossed with an avirulent (LD50 = 108)S.montevideo donor (Hfr) strain and his+
recombinants were selected. As expected (Makela, 1966), about 80 % were serologically 6, 7 and had therefore inherited the ifbc cluster from the S. montevideo donor,
while the remaining 20 % were 4, 5, 12 and had thus retained the original rfbB locus.
Virulence of two his+ 6,7 and two his+ 4 , 5 1 2 recombinants was determined (Table 6).
These recombinants were also trp- like the recipient, which put a limit on the amount
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Thu, 13 Jul 2017 00:41:57
264
V. V. V A L T O N E N
of donor material on one side of the his marker (see Fig. 2 ) ; on the other side the 6 , 7
recombinants could have inherited a larger fraction of the donor chromosome. This
fraction was probably quite small, however, and the two recombinant classes
differed from each other mainly in the rj‘b region, because in intergroup crosses only
short chromosome fragments are usually incorporated (Stocker & Makela, 197I).
The two recombinants with 4,5, 12 specificity were as virulent as the recipient (LD50
105),while the 6,7 recombinants were avirulent (LD50 = 10’ to 1 0 ~ )The
.
difference
between the LD50 values of these two recombinant classes was highly significant
(P < 0.001).As a further check it was tested whether the virulence of the 6 , 7 recombinant could be restored to the level of the parent, by reintroducing the rfbB locus
determining 4, 5, 12 specificity, to a his- mutant of the 6, 7 recombinant. This ‘backcross’ was performed using the same s.abony Hfr donor as with the 9,12transductants
above. The LD50 values of two his+ 4, 5, 12 recombinants were between 105 to I O ~ ,
while the LD50 value of a his+ 6, 7 recombinant was unchanged (IO~), strongly
indicating that the 4, 5, 12 side-chains had restored virulence to these recombinants
(Table 6). The fact that the original LD50 value of 105 was not quite achieved could
be due to the manipulations (mutagenesis and crossing), which the strains had
experienced before virulence testing.
-
Table 6. LD50 values of smooth Salmonella typhimurium strains
with the 0 specijicity 4, 5, 12 or 6, 7
Origin of recombinants
or transductants
I
A
\
Donor
Recipient
Cross (conjugation)
I
L
~~1605x
\
~ ~ 2 1 8 3
LD 50 values
h
f
Strain no.
Recipient
~ ~ 2 1 8 3
Recombinants
SH 2245
SH 2246
SH 2241
SH 2242
Back-cross (conjugation)
Recipient
SH 2259
Recombinants
Selection for his+
P22 transduction
SH 2273
SH 2274
SH 2282
Recipient
SH I
SH I01I
X
(P22,his+, 6, 7)
SHI
(his-, 4,5,12)
Selection for his+
Transductants
SH 1331
SH I333
~~1360
SH 1361
\
0 specificity
A
I
495, 12
I x 105
6, 7
I x 105
2 x 106
-
-
>
-
4 X 107
Ix
106
-
n.d.
5 x 106
8 x 106
-
-
1 x I08
4 x 106
-
2 x 105
8 x 106
-
-
-
n.d. = not determined.
~~225
(his-,
9 6,7) was derived from recombinant ~ ~ 2 2 4 1 .
-
Ix
Ix
107
108
It seems very probable, therefore, that the loss of virulence in the 6, 7 recombinants
was due to the presence of the rfbc locus, although possible ‘avirulence loci’
(Krishnapillai & Baron, 1964) closely linked to his and present in the avirulent
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Thu, 13 Jul 2017 00:41:57
Virulence of Salmonella and 0 antigen
265
Salmonella montevideo donor, could not be excluded. Because of the drastic reduction
of virulence it was felt desirable to limit the amount of genetic material transferred
along with his. This could be achieved in transduction, but the available transducing
phage P22 would not attack 6, 7 bacteria. A P22 lysate could, however, be prepared
by induction of a strain of S. typhimurium which was lysogenic for P22 and which had
been given ifbc in a conjugational cross with S. montevideo as the Hfr donor. The
8 a3 recilysate had a fairly low titre, and transduction did not succeed with ~ ~ 2 1 as
pient. It did succeed, however, with another his- S. typhimurium strain LT2 (SHI).
The LD50 values of two his+ 4,5,12 transductants were again 105 to log like those
of the recipient, while the corresponding values of his+ 6, 7 transductants were 10'
to IO* (see Table 6).
DISCUSSION
This work has demonstrated that different 0 side-chains confer different degrees of
mouse virulence on a basically virulent Salmonella typhimurium. Salmonella typhimurium with the original 4, 5,12-specific side-chains is slightly more virulent than the
same strain with 9, 12 specificity, and this is clearly more virulent than the same
S. typhimurium strain with 6, 7 side-chains. Yet all the strains studied were smooth
according to all available criteria : sensitivity to S-specific phages and resistance to
R-specific phages ; S-type agglutination with specific anti-0 sera but no agglutination
with 4 % (w/v) saline or with antirough sera.
In each experiment the virulence of several sister recombinants was compared in
order to keep the strains as isogenic as possible, the only difference between them
being in the 0 side-chain. When comparing the 9, 12 and 4, 5, 12 strains this purpose
appears to have been well achieved. The strains were prepared by transduction, so
that the non-identical fraction of the genome of the two transductant classes was
less than I % (the amount of DNA transferred in transduction is of the order of I %
of the total chromosome; both transductant classes also had common areas in the
histidine locus received from the donor). The 6, 7 strains were prepared by conjugation and by transduction. In both cases the result was the same, low virulence always
accompanying the 6, 7 characteristic and high virulence accompanying the 4, 5, 12
specificity. Combining the data from the transduction and the conjugation analyses,
four strains of 9, 12 specificity were all less virulent than six sister 4, 5, 12 strains and
five strains of 6, 7 specificity were clearly less virulent than seven sister 4,5,12 strains.
There is strong evidence in these results for the association of virulence with the
quality of the 0 side-chain as determined by the rfb locus. Other genes determining
virulence and closely linked with the rfb locus are not excluded but if they exist their
number is limited.
A limitation of the action of foreign (groups C or D) genetic material in the group B
genome is also possible. Again the effect, if present, must be limited to rfb and a few
other closely linked genes. If this impairment occurred, it was not large enough to
make the strains detectably non-smooth, although searching criteria for smoothness
were applied (Lindberg, Sarvas & Makela, 1970). Pleiotropic effects of the foreign rfb
genes are also conceivable but improbable.
We do not know how the quality of the 0 side-chain affects virulence. There are at
least two possibilities: (a) the quality of the monosaccharide constituents of the
different types of 0 side-chains is of major importance, or (6) the main factor is the
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Thu, 13 Jul 2017 00:41:57
266
V. V. V A L T O N E N
degree of exposure of the deeper layers of the cell wall including the LPS core, resulting from differences in length and/or arrangement of the 0 side-chains of different
specificities.
The quality of the monosaccharides is very similar in the repeating units of the 4 1 2
and 9, 12 side-chains. The only difference lies in the dideoxyhexose which forms an
immunodominant side branch in the main chain (Fig. I), and which is abequose in
4,12 and tyvelose in 9,12. It is not easy to understand why the mouse reacts differently
to the two sorts of dideoxyhexoses. The 4,12 strains used in this work also had factor 5,
which corresponds to an 0-acetyl group on the abequose (Hellerqvist et al. 1968),
but in another study isogenic Salmonella typhimurium strains with and without factor 5
were found to be equally virulent, indicating that this factor does not affect virulence
(V. V. Valtonen & P. H. Makela, in preparation). The 6 , 7 side-chains on the other
hand are very different, lacking both the deoxyhexoserhamnose and the dideoxyhexose.
From the point of view of the host in relation to the observed differences in virulence, two mechanisms can be suggested: resistance could be based either on an
immunological mechanism or associated with phagocytosis. The 6, ppecific sidechains may be more immunogenic than the 4,5,12 side-chains, or the mice may have
‘natural antibodies ’ against 6, 7 antigen giving better protection against a Salmonella
strain with 6, 7 specificity. If these were the main causes of the observed differences
in virulence, we have to assume that a 9, 12-specific LPS is also more immunogenic
than the 4, (9,12 LPS, or that the mice possess more ‘natural antibodies’ against
9,12 than 4, (5),1z determinants. We have studied the role of the immune response in
infection with the transductant strains used in this work and the results are to be
published soon.
There are several reports which have shown that avirulent Salmonella strains are
killed more rapidly by phagocytic cells than are virulent ones (Furness, 1958; Biozzi
et al. 1964; Fauve, 1964; Ushiba, 1965). The mechanism by which virulent strains
can resist phagocytosis is unknown, but it seems probable that the structure of the
LPS is of major importance in this respect. An obvious experimental approach would
be to determine whether the 4, 5, 12 LPS is more ‘antiphagocytic’ than 9, 12 or 6 , 7
LPS.
Similar differences in virulence associated with cell-wall polysaccharides have also
been found in Pneumococcus and Escherichia coli. MacLeod & Krauss (1950) found
that both the amount and the quality of the capsular polysaccharide have an effect on
the virulence of Pneumococcus. The capsular polysaccharides are known to be ‘antiphagocytic’ substances. The capsular polysaccharide of type VII Pneumococcus is
a particularly efficient antiphagocytic substance ; it contains L-rhamnose and other
deoxysugars (How, Brimacombe & Stacey, 1964). Medearis and co-workers (1968)
have found that the sugar composition of the cell wall strongly affects the virulence of
E. coli. Apart from a rough mutant they also used an apparently smooth mutant of
E. coli 0 - 1 1 1 strain lacking only a terminal dideoxyhexose colitose in its LPS, and
found this to be clearly less virulent than the parent strain. Loss of virulence in this
mutant was accompanied by diminished resistance to phagocytosis. The present work
with Salmonella strains has clear analogies with these observations. The 4, 5, 12 and
9, 12 side-chains, which confer stronger mouse virulence than the 6, 7 side-chains,
also contain both L-rhamnose and a dideoxyhexose (abequose or tyvelose respectively), while the 6 , 7 side-chain does not contain either. One is tempted to suggest that
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Thu, 13 Jul 2017 00:41:57
Virulence of Salmonella and 0 antigen
267
these lipophilic deoxy- or dideoxysugars,which are rare in nature, are also important
determinants of anti-phagocytic activity and thus of bacterial virulence.
I thank Professor P. Helena Makela for her help in preparing the manuscript and
also for help with materials, Professor 0. Miettinen for suggesting the statistical
significance test applied, Miss Marianne Hovi for preparing some of the strains, and
Mrs Aila Akerlund and Mrs Eila Tikkanen for their technical assistance. This work
was supported by grants from the Sigrid Jus6lius Foundation to Professor Makela and
from the Emil Aaltonen Foundation.
REFERENCES
ARCHER,
J. R. & ROWLEY,
D. (1969).A quantitativecomparisonof the antigenic structure of a virulent
and an avirulent strain of Salmonella typhimurium. Immunology 17,551-558.
ARKWRIGHT,
J. A. (1927). The value of different kinds of antigen in prophylactic 'enteric' vaccines.
Journal of Pathology and Bacteriology 30, 345-364.
BIOZZI,G., LE MINOR,L., STIFFEL,
C., MOUTON,
D. & DAOULAS,F. (1964). Correlation between
virulence and phagocytosis of genetic recombinant between Escherichia coli and Salmonella
typhimurium. Nature, London 202, 819-820.
EDEBO,L. & NORMA", B. (1970). Virulence and immunogenicity of mutant strains of Salmonella
typhimurium. Acta pathologica et microbiologica scandinavica 78 B, 75-84.
FAUVE,M. R. (1964). Rksistance cellulaire l'infection bactbrienne. Annales de l'lnstitut Pasteur,
Paris 107,472-483.
FULLER,
N. A. & STAUB,
A. M. (1968). Immunochemicalstudies on Salmonella 13. Chemical changes
appearing on the specific polysaccharideof S. cholerae suis (&, 7) after its conversion by phage 14
(6, 7). European Journal of Biochemistry 4, 286-300.
FURNESS,
G. (I 958). Interaction between Salmonella typhimurium and phagocytic cells in cell culture.
Journal of Infectious Diseases 103,272-277.
GEMSKI,
P. JUN. & STOCKER,
B. A. D. (1967). Transduction by bacteriophage P 22 in non-smooth
mutants of Salmonella typhimurium. Journal of Bacteriology 93, 1588-1 597.
HARTMAN,
P. E., LOPER,J. C. & SERMAN,
D. (1960). Fine structure mapping by complete transduction between histidine-requiring Salmonella mutants. Journal of General Microbiology 22,
323-35 3.
HELLERQVIST,
C. G., LINDBERG,
B., SVENSSON,
S., HOLME,
T. & LINDBERG,
A. A. (1968). Structural
studies on the 0-specific side-chains of the cell wall lipopolysaccharide from Salmonella
typhimurium 395 MS.Carbohydrate Research 8,43-55.
HERZBERG,
M. (I 962). Living organisms as immunizing agents against experimental salmonellosisin
mice. I. Virulence of auxotrophic mutants. Journal of Infectious Diseases 111, 192-203.
How, M. J., BRIMACOMBE,
J. S. & STACEY,
M. (1964). The pneumococcal polysaccharides. Advances
in Carbohydrate Research 19,303-358.
KAUFFMANN,
F. (I966). The Bacteriology of Enterobacteriaceae. Copenhagen: Munksgaard.
KRISHNAPILLAI,
V. & BARON,L. S. (1964). Alterations in the mouse virulence of Salmonella typhimurium by genetic recombination. Journal of Bacteriology 87,598-605.
LEDERBERG,
J. (I 950). Isolation and characterization of biochemical mutants of bacteria. Methods in
Medical Research 3, 5-22.
LINDBERG,
A. A., SARVAS,
M. & MAKELA,P. H. (1970). Bacteriophage attachment to the somatic
antigen of Salmonella: effect of 0-specific structures in leaky R mutants and S, T I hybrids.
Infection and Immunity I, 88-97.
LINGELSHEIM,
V. (1913). Zur Frage der Variation der Typhusbacillen und vmandter Gruppen.
Zentralblatt fur Bakteriologie, Parasitenkunde, Infektionskrankheiten und Hygiene (Abteilung I
Originale) 68, 577-582.
LOPER,J. C., GRABNAR,
M., STAHL,R. C., HARTMAN,
Z. & HARTMAN,P. E. (1964). Genes and proteins involved in histidine biosynthesis in Salmonella.Brookhaven Symposia in Biology 17,I 5-52.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Thu, 13 Jul 2017 00:41:57
268
V. V. VALTONEN
LUDERITZ,O., JANN,K. & WHEAT,R. (1968). Somatic and capsular antigens of Gram-negative
bacteria. In Comprehensive Biochemistry, vol. 26 A, pp. 105-228. Edited by M. Florkin and
E. H. Stotz. Amsterdam: Elsevier.
MACLEOD,
C. M. & KRAUSS,
M. R. (1950). Relation of virulence of pneumococcal strains for mice
to the quantity of capsular polysaccharide formed in vitro. Journal of Experimental Medicine
92, 1-9.
MANTEL,
N. & HAENSZEL,
W. (1959). Statistical aspects of the analysis of data from retrospective
studies of disease. Journal of The National Cancer Institute 22, 719-748.
MEDEARIS,
D. N. JuN., CAMITTA,
B. M. & HEATH,E. C. (1968).Cell wall composition and virulence
in Escherichia coli. Journal of Experimental Medicine 128, 399-414.
MAKELA,P. H. (1963). Hfr males in Salmonella abony. Genetics 48, 423-429.
MAKELA,P. H. (1966). Genetic determination of the 0 antigens of Salmonella groups B (4,5,12) and
C1 (6, 7 ) .Journal of Bacteriology 91,I I 15-1 125.
MAKELA,P. H. & STOCKER,
B. A. D. (1969). Genetics of polysaccharide biosynthesis. Annual Review
of Genetics 3, 291-322.
NAKANO,
M. & SAITO,K. (1969). Chemical components in the cell wall of Salmonella typhimurium
affecting its virulence and immunogenicity in mice. Nature, London 222, 1085-1086.
REED,L. G. & MUENCH,
H. (1938).A simple method of estimating fifty per cent end-points. American
Journal of Hygiene 27,493-497.
ROANTREE,
R. J. (1967). Salmonella 0 antigens and virulence. Annual Review of Microbiology 21,
443-466.
SANDERSON,
K. E. (1 967). Revised linkage map of Salmonella typhimurium. Bacteriological Reviews
31,354-372.
STOCKER,
B. A. D. & MAKELA,P. H. (1971). Genetic aspects of biosynthesis and structure of Salmonella lipopolysaccharide. In Microbial Toxins, vol. 4. Edited by S. J. Ajl. New York: Academic
Press.
USHIBA,
D. (1965).Two types of immunity in experimental typhoid: ‘cellular immunity’ and ‘humoral
immunity’. Keio Journal of Medicine 14, 45-61.
VALTONEN,
V. V. (1969).Virulence of Salmonella strains with a reduced amount of 0 antigen. Journal
of General Microbiology 57, xxviii.
WILKINSON,
R. G. (1966). Rough Mutants of Salmonella typhimurium. Ph.D. Thesis, University of
London.
B. A. D. (1968). Genetics and cultural properties of mutants of
WILKINSON,
R. G. & STOCKER,
Salmonella typhimurium lacking glucosyl or galactosyl lipopolysaccharide transferases. Nature,
London 217, 955-957.
Downloaded from www.microbiologyresearch.org by
IP: 88.99.165.207
On: Thu, 13 Jul 2017 00:41:57
.